FSR-LPTV filtering for blind equalization and interference rejection

We address the problem of synthesizing a blind channel identification and equalization method for digital communications systems, aimed at counteracting also the presence of co-channel or adjacent-channel interference. Owing to the presence of the interfering signal, the minimum mean-square error equalizer turns out to be linear periodically time-varying (LPTV), which is implemented by restoring to its Fourier series representation (FSR). Moreover, by exploiting the cyclic conjugate second-order statistics of the channel output, we propose a new weighted subspace-based channel identification method. Computer simulation results confirm the effectiveness of the proposed technique.